DeepSeek V3.2 Speciale
DeepSeek-V3.2-Speciale is a high-compute variant of DeepSeek-V3.2 optimized for maximum reasoning and agentic performance. It builds on DeepSeek Sparse Attention (DSA) for efficient long-context processing, then scales post-training reinforcement learning to push capability beyond the base model. Reported evaluations place Speciale ahead of GPT-5 on difficult reasoning workloads, with proficiency comparable to Gemini-3.0-Pro, while retaining strong coding and tool-use reliability. Like V3.2, it benefits from a large-scale agentic task synthesis pipeline that improves compliance and generalization in interactive environments.
- Input / output price
- $0.58 / $1.68 per 1M
- Context
- 128K tokens
- Released
- Dec 1, 2025
Paste into Claude Code, Cursor, or Codex
Help me set up the Respan gateway: fetch and follow https:/
Providers
One provider serves this model, priced per 1M tokens after gateway discounts.
| Provider | Input | Output |
|---|---|---|
Azure Foundry azure_foundry/deepseek-v3.2-speciale | $0.58/M | $1.68/M |
Pricing
USD per 1M tokens on Azure Foundry, after gateway discounts.
- Input
- $0.58 per 1M
- Output
- $1.68 per 1M
Cost by volume
Total cost from 0 to 1T tokens.
Benchmarks
10 published scores from 4 sources, each linked to where it was published.
Published scores
| Benchmark | Score | Source | Published |
|---|---|---|---|
| LiveCodeBench (Pass@1-COT)huggingface.co | 88.7% | huggingface.co | Dec 1, 2025 |
| IMOAnswerBenchhuggingface.co | 84.5% | huggingface.co | Dec 1, 2025 |
| WeirdMLHåvard Tveit Ihle | 46.7% | Håvard Tveit Ihle | Oct 5, 2026 |
| Benchmark | Score | Source | Published |
|---|---|---|---|
| GPQA Diamondhuggingface.co | 85.7% | huggingface.co | Dec 1, 2025 |
| SimpleBenchSimpleBench | 52.6% | SimpleBench | Dec 12, 2025 |
| HLE (text-only)huggingface.co | 30.6% | huggingface.co | Dec 1, 2025 |
| Benchmark | Score | Source | Published |
|---|---|---|---|
| HMMT Feb 2025huggingface.co | 99.2% | huggingface.co | Dec 1, 2025 |
| AIME 2025huggingface.co | 96% | huggingface.co | Dec 1, 2025 |
| AIME 2025MathArena | 95.8% | MathArena | Oct 13, 2026 |
| HMMT Nov 2025huggingface.co | 94.4% | huggingface.co | Dec 1, 2025 |
Apps
Use DeepSeek V3.2 Speciale in the tools you already run. Each guide points the app at the Respan gateway.
Setup guides for more apps are in the integration docs.
DeepSeek V3.2 Speciale FAQ
Short answers, built from the data on this page.
DeepSeek V3.2 Speciale is a model from Deepseek, released on Dec 1, 2025. DeepSeek-V3.2-Speciale is a high-compute variant of DeepSeek-V3.2 optimized for maximum reasoning and agentic performance. It builds on DeepSeek Sparse Attention (DSA) for efficient long-context processing, then scales post-training reinforcement learning to push capability beyond the base model. Reported evaluations place Speciale ahead of GPT-5 on difficult reasoning workloads, with proficiency comparable to Gemini-3.0-Pro, while retaining strong coding and tool-use reliability. Like V3.2, it benefits from a large-scale agentic task synthesis pipeline that improves compliance and generalization in interactive environments.
On the Respan gateway, DeepSeek V3.2 Speciale costs $0.58 per 1M input tokens and $1.68 per 1M output tokens on Azure Foundry, after gateway discounts.
DeepSeek V3.2 Speciale is served by Azure Foundry.
DeepSeek V3.2 Speciale accepts up to 128K tokens of context.
Point any OpenAI-compatible SDK at https://gateway.respan.ai/v1 with your Respan API key and set the model to "deepseek-v3.2-speciale". To set it up with a coding agent, copy the agent setup prompt at the top of this page.
DeepSeek V3.2 Speciale has 10 published benchmark scores, including LiveCodeBench (Pass@1-COT) 88.7%, GPQA Diamond 85.7%, and HMMT Feb 2025 99.2%. Each score links to where it was published; the data is collected by Respan.
Certified where it counts.
Built to meet the security and privacy standards that enterprise and healthcare teams require.
ISO 27001
The internationally recognized standard for information security management.
SOC 2
Secure, compliant management of your data across all of our systems.
GDPR
Operated under GDPR, the world's strictest standard for data privacy.
HIPAA
HIPAA compliant, with a BAA available for healthcare teams.